[randytsuch]

Apparently, the 3.3Vdd for the DAC is not coming from a separate regulator off the 14Vdc rails, but is coming off the switching regulator (the 1st regulator shown in the above diagram) which is also powering a slew of other stuff. Moving it to the linear regulator powering the headphone amp, assuming it is of sufficient quality and quantity, might be a good idea.

If you did this, I would consider removing the headphone amp, if you are not using it.  I have seen designs with a regulator for every device.

Randy

[Occam]

Some thoughts on quick n' dirty mods to the SB2 -

The easiest is replacing the switching wall wart with a better perfoming linear supply. I loaned Josh a linear brick, and found him a source for standard industrial 5v 3a linear, filtered supplies. The $8 brick is substantially more than the $3.50 3a linear, but the brick already has a cord, enclosure..... After Josh has given these a breif evaluation, he'll post the source for the preferred supply.

I've considered the potential benefits balancing/powerconditioning (Felicia, or a simplified specific version) but don't think it really worth the effort. Not that this might not provide substatial benefits, but simply for the fact that this really is an application where a battery would be just swell. No 12v batteries are needed, a 6v SLA will do just fine. A large benefit is for the same size and weight as a 5a 12v SLA, you get a 12a 6v SLA. (which means for the same nominal use, when we responsibly recycle our dead batteries, we only ship half the lead to China to pollute their environment).

So you get a 6v 12a SLA-
http://www.gruberpower.com/Products/Batteries/VoltAmp/6Volts/
and a cheap (but competent) charger -
http://www.zbattery.com/6bc1000d-1.html

and now you want to drop that 6.3-6.4 volts down to that nominal 5vdc. So you wire up a high spec low noise, primo for audio use, low drop out regulator, a LT1764ET (or possibly a LT1963AET)-
http://www.linear.com/pc/productDetail.do?navId=H0,C3,P2221
http://www.linear.com/pc/productDetail.do?navId=H0,C3,P2222

Although straightforward, I've no regulator boards to offer, or point anyone to, as of yet.
[a note to Vendors - as moderator, I've no objection to vendor participation, and if someone would like to offer regulator boards, kits, etc.... on the Lab Circle, I encourage it, with the only previso being that any such offerings must be open source, you must disclose the schematic so if a participant doesn't wish to purchase your product, they can build from the schematic]

It might be appropriate to draw upon life experience here..... Sometimes its rather enjoyable to be passive, rather than active .  Not allways the best, but sometimes quite refreshing and interesting. Sooooo, you might consider dropping the voltage with an appropriate diode, with a diode drop, Vf, of 1.1-1.4volts. Bypass the battery with 4K+-uf of 10+v electrolytics of your choice, add the diode, and bypass again. Bypass the bypasses as your religion dictates, .1uf relcaps, etc....
Not the greatest regulation, but whatever  noise from the battery and diode is quite amenable to that passive treatment.

The above is really quite straightforward as a simple ps upgrade. But what can we do about the ps on the output section, if you want to improve the analog out. No matter what we do, unless we wish to disable the display, that switcher/multipier is going to be active. This switcher/multiplier also powers the 78xx type 5v Vacc for the DAC and the 9v single rail for the output opamps. But if we accept the philosophy of Thorsten/Yeo/Vinnie/Wayne et al, we can live with the constraint of the limited output swing and drive, and take our analog outputs directly from DAC, via the cap of your choice. Now, we don't need to drive a 9v regulator and can put an extrememly low noise 5v regulator for the DACs Va (and a 3.3v for the Vd if we can't use the headphones 3.3v regulator); something like a LT1761 -
http://www.linear.com/pc/productDetail.do?navId=H0,C1,C1003,C1040,C1055,P1750
and feed it with the 6volts from the SLA battery and eliminate the potentially noisey feed from the switcher/multiplier.

Eliminating the output NJM2401 opamps still leaves you with an audio output of 3.1v p-p, which is a little over 2Vrms.

If you're content with a new 5v supply feeding the standard power socket, you might consider upgrading the opamps to either an opa2107 or ad823 which in S08 should drop in, and might be subjectively prefferable. (the ad823 isn't my cuppa)

And while your hacking up your SB2 and voiding the warranty (unless your willing to content yourself with the upgraded 5v supply feeding the standard power socket, which for those who want to feed an external dac, might be quite good), you might want to hardwire interconnects terminated in RCAs, and a SPDIF terminatied as appropriate, RCA, BNC,SMC... eliminating some connections.
FWIW

[JoshK]

You da man!

[Occam]

You da man!

Oh, how I yearn to hear those words again...... from my wife  

Josh and I discussed much of my prior post above last night, and he suggested someone who might be willing to provide regulator boards/kits. I will pursue this option. And thankfully, Josh has decided to undertake this endeavor with something approaching rigor. PhilNYC is providing his SB2 as the baseline machine, so Josh actually has something to compare the modded SB2 against, rather that relying on notoriously questionable aural memory. As mods are made and vetted, only then will the baseline machine be upgraded.

Dang! If this thing goes well, I might even buy a SB2.

[randytsuch]

Here is a regulator for you
http://www.aoselectronics.com/jungsrpcb.html

It is only the PWB, but it is cheap, and it is supposed to be the best regulator you can get for audio.  The problem is he only has a schematic, so you would need to create a parts list, and figure out what to buy.

There is a guy in England that sells kits for jung regulators, but they are much more expensive.  I built a couple of them, but have not completed the project they are used in.

Randy

[Davey]

Paul,

3.1v p-p is 1.09Vrms.  (That might be less voltage than some folks require.)

On my totally unmodded unit the analog outputs measure 1.86Vrms playing a 0dbFS signal.

Cheers,

Davey.

[Occam]

Davey - Yea, but I was only off by a factor of 2.... I'm usually off by a factor of 10  
Mea culpa.....

[JoshK]

I forget the rationale behind why regulation is not a critical issue.  Can you explain that a little?

[Occam]

Josh - the rational behind my less than obscessive concern over line and load regulation is the nature of the power draw on the on the SB2, its pretty much constant. The 5v enters the SB2 and immediately is processed by further voltage regulation which largely determines its load and line regulation peformance. As long as the voltage/current is sufficient to power these follow on regulators, it should work well.
If you look at the tripath based, battery powered amps, you'll note that they typically do not have regulators further processing the power, but rely upon the nature of the batteries themselves to provide whatever line and load regulation that is required.
Noise is another matter, and should be adressed as early in the powertrain as possible. It propogates, and even within a fully regulated environment, it is largely addressed by the passive components (capacitors) that are a part of whatever regulation circuitry.
That being said, the 2 previously mentioned regulators are probably preferable to the simple diode drop previosly discussed, if only for their ability to provide regulated power with a maximum of a .34volt dropout. Nor is the additional circuitry overly onerous. The diode requires a capacitor before and after, where substituting a monolithic regulator needs similar flanking capacitors and would only require an additional pair of resistors for the voltage divider (these are adjustable regulators) and possibly an additional  cap to bypass the ground leg of the resistive divider.
No doubt, someone could design a regulator with better noise and regulation characteristics that the Linear regulators mentioned with discrete circuitry, but I'd have no idea as to how to do it with such a low dropout. Remember, we are trying to use a nominal 6v battery to provide a regulated 5Vdc. The very high performance Jung/Didden reguators  require a minimum of 2volts+ headroom to work.

[Davey]

Paul,

Here's a question for you.  Why do you have to drop the voltage from the battery to 5 volts?  Why not just hookup the battery and forget about it?  The SB2 internal regulators shouldn't care if their input voltage is a bit higher eh?

Cheers,

Davey.

[Occam]

Davey,

Excellent question and idea! The answer is - dunno. The input switcher might well be able to handle the higher voltage with aplomb, via its self regulating nature, maybe not.... That input 3.3v linear regulator would have to dissapate substantilly greater power. Even if you disable the headphone amp which it powers, I'd try to use it to power the DAC's digital circuitry, which is presently powered by that input switcher, about 100 x [(6.3-3.3)/(5-3.3)-1] or  76% more, but it might be feasable with a glued on heatsink.... Then again it might not be problematic if you can forgo the headphone drive capability and disconnect the V+ lead to the headphone amp, in that it wouldn't be powering a headphone amp driving a 32?ohm load, but the digital V for the DAC which is a maximum of 10ma+-.
So my answer would be, it depends. I see no heatsinking on the regulators I've seen in pictures (I don't have a SB2), nor do I know the overvoltage tolerances of the input switcher. Your query leads to some very interesting extentions. That 'raw' battery input could also feed higher spec regulators for the output circuitry such as a LT1761/2, 3.3v for the Dac's digital and 5v for its analog. While directly driving the output from the DAC's internal opamps has appeal, its 50pf load limitation is of concern. A typical RCA plug/socket mate has minimum of 20pf+ alone.  With 17pf/ft for a low capacitance coax like 89259 for analog interconnects, we rapidly run into constraints without even considering the load capacitance.  (though a 10-22ohm series load might decouple the load somewhat). This is among the reasons why I suggested hardwiring the analog outputs. But if we feed the output circuitry with the raw battery power of 6.3-6.4Vdc and regulate down, we could directly couple the dac to a single rail ad8066, fed by another 5v reguator. (or raise the analog output circuitry rail to 5.5v for a slight increase in swing). While this isn't going to raise the output swing to 2Vrms, it will make the SB2 quite a bit more flexible in what it can drive. Certainly, eliminating the external opamps has a theoretical (and possibly a real) advantage, my preferred analog coax interconnect is 30pf/ft.....

[JoshK]

I love this sort of discourse!     I learn a lot from reading these kinds of posts and it inspires me to go look up specs on said opamps to see if I can see the conclusions drawn.  Please, keep it coming.

[Davey]

Well Paul, you're doing a lot of re-designing there.    Currently, I'm happy to use mine just the way it is.  If you didn't have the other usage in mind I'd just clip the input lead on the 3.3v regulator.  No more possible dissipation problem.  

Regarding the capacitive loading limitation on the DAC outputs.....I would have thought Vinnie and some of the other modders would have already looked into this and found the situation acceptable.  He doesn't mention adding a de-coupling build-out resistor in his mod so I dunno if he did or not.

I always find these type of mods a bit puzzling.  (But that's just me, I'm an oddball.)
Usually it seems like the modder is just throwing the standard-audiophile-mod-playbook at the "problem."  Op-amps are terrible...let's get rid of them.  Switcher convertors are terrible...let's get rid of those too.  Batteries are so "pure"...let's use those.  Let's put some really high-dollar RCA jacks in there too.  And there's never any objective testing published to justify it.  It's all about the "S."  Subjective evaluation.  You see what I mean.

I might pop the hood on my SB2 one of these days but currently I find fiddling with speakers much more interesting/enjoyable.

Cheers,

Davey.

[Occam]

Well Paul, you're doing a lot of re-designing there.  

Thats why this forum is called the Lab as opposed to the Kludge forum.....

If you didn't have the other usage in mind I'd just clip the input lead on the 3.3v regulator. No more possible dissipation problem.

And you've verified that that input switcher can handle the direct battery voltage? But I do have another possible use in mind. Regardless, a battery, monolithic regulator, 3 caps, and 2 resistors doesn't strike me as going round the bend. Given my ongoing affair with 'Felicia', I don't think anyone would accuse me of being 'battery mad'. This really is a situation where a 6v SLA, recharger, and a simple regulator is the simplest solution to that sucky switching wall wart. And for those using an external DAC, with the SB2's impressive jitter performance, you're almost home. (even if you don't intend to put in a digital output transformer, at least hardwire a properly terminated 75ohm cable to the S/PDIF coax out, or heavens forfend, try a GLASS toslink cable)

Regarding the capacitive loading limitation on the DAC outputs.....I would have thought Vinnie and some of the other modders would have already looked into this and found the situation acceptable. He doesn't mention adding a de-coupling build-out resistor in his mod so I dunno if he did or not.

I've no issues with Vinnie's (or Wayne's) mods. I've heard Vinnie's stuff and been mightily impressed. But he does suggest short interconnects. Besides, show me someone with a system that measures, in-room at the listening position, flat to 20kz, and I'll show you someone who doesn't listen to their system very often.

Regarding the rest.... Life is full of compromises. JoshK's goal is to get maximum bang for the buck. DIYing is quite different from making marketable mods. PhilNYC had donated his SB2 as a baseline system so that comparisons can be made. I'm certainly not advocating putting in bespoke RCA sockets, quite the opposite. And yes, I'd ideally like to eliminate switchers powering the output circuitry, because I know noise using such regulators is substantially higher. And while I'd ideally like to remove all opamps (besides those internal to the DAC), I actually suggested an AD8066, which I've heard and subjectively like very much, which would provide more drive, run off of 5v single rail, and eliminate the electrolytic coupling cap between the DAC and opamps.

Its great that you like your SB2 as is. If you think your time and resources are better spent on your speakers rather than replacing the 5v wallwart switcher with a linear or battery ps, so much the better.

[JoshK]

But hacking apart gear for incremental gains is so much fun!  

Seriously though, it is an interesting project, IMO should contribute modest gains and doesn't seem too hard or expensive.  Why shoud we let the pro-modders have all the fun?  

[Davey]

Josh/Paul,

Yes, indeed...good points.  I will bow out of this thread until I can contribute something useful.  

Cheers,

Davey.

[Occam]

Davey,

Why in heaven's name bow out as your posts ARE extremely useful?
Your proposing runninng directly from a SLA raised issues about dissapation limits of the linear 3.3v regulator, tolerances of the initial switcher, and raised the possibility of powering directly the output stage regulators. IMO, very, very helpful. (additionally, I allways need someone to check my math   ) Please continue to post; we need your input.
Nor does my initially disagreeing with your approach mean that you won't ultimately be proven correct.

Regards,
Paul

[BobM]

But hacking apart gear for incremental gains is so much fun!   quote]

Spoken like a true solder addict. Watch those fumes now guys!
 

I'd be sorely tempted to join in on the fun except for the following:
- I really don't want to burn my thousands of CD's to my PC hard drive
- I'm playing with vinyl more these days (new tonearm)
- I'm salivating to hear the feedback on the new Hagerman DAC that's forthcoming
- It's summer and I'm playing OUTDOORS

Enjoy,
Bob

[JoshK]

I am not planning to use the internal DAC in the SB2 except to help exact mods.  I am planning mostly to mod the power and digital transmission sections of my SB2.

[JoshK]

I am going to add a poll to the initial post so I can gauge how many people are interested in digital output only, analog output only, both, neither.  

For those who are interested in the digital output side of things, I am really interested in sourcing digital output transformers.  www.scientificonversion.com is one such source.  (Aberdeen Components gave me that source, thanks!).  

It looks like they take small orders but maybe it would be a lot nicer if someone (like me or other volunteer) organizes a group purchase instead, so they aren't annoyed with a bunch of orders for one transformer.  We, after all, want these types of manufacturers to keep participating with DIYers.   If we get a discount for larger purchase, all the better, but I wouldn't hold my breathe that we would get sufficient qty for discount.

So please email/pm me if you are interested in ordering a transformer and when I get a ballpark number I'll email them for prices.

The transformers in question are:
SC947-02

These are SMD transformers, which may be a pain for some to work with so if I can find info that says their thru-hole trannies work as easily, then we could consider those instead.

Myself, I am super curious about the Dual Zo 75 -> 75/110 transformer as I'd love to use the AES input on my DEQX instead of the coax.  My coax I'd prefer to use otherwise, this transformer would allow both AES & coax output but an appropriate hack to the case for an XLR would be needed.